Hi, I think this post will get much more attention on the Red Hat rpm-list. Matthias Zhou Wei wrote : > Hello, > > I have some questions about device number extension. > > In Linux kernel 2.6, device number will be extended from 16-bit to > 32-bit. All utilities and libraries should make corresponding extension > for this new feature in kernel 2.6. > > I find that __rpm-4.2.1-4.2__ uses structure dev_t and operates the > device number as > 16-bit.----------------------------------------------------------------- > ------------------------- 1. system.h line 76 and file/system.h line 165 > :#ifndef HAVE_MAJOR > #define major(dev) (((dev) >> 8) & 0xff) > #define minor(dev) ((dev) & 0xff) > #define makedev(maj, min) (((maj) << 8) | (min)) > #endif > > This is a backup for lost of system defined macro, but all of these > operations look dev_t as 16 bits instead of 32 bits, so they will call > problem in current > situation.-------------------------------------------------------------- > ---------------------------- > > 2. build/files.c line 485 and 501 > if (!(fl->devmajor >= 0 && fl->devmajor < 256)) > if (!(fl->devminor >= 0 && fl->devminor < 256)) > > This sentence confirm the range of major device number and minor device > number. But major device is 12bits and minor device is 20 bits now, so > their max value isn't 255 any > more.------------------------------------------------------------------- > ----------------------------------- > > 3. build/files.c line 1234 > if (sizeof(flp->fl_rdev) != sizeof(uint_16)) > uint_16 prdev = (uint_16)flp->fl_rdev; > > flp->fl_rdev is dev_t type and it is 32 bits, so this assign 32bits to 16 > bits, it may cause problem. > ------------------------------------------------------------------------ > ------------------------------ > > 4. lib/verify.c line 180 > uint_16 st_rdev = (sb.st_rdev & 0xffff); > > sb is a stat struct and sb.st_rdev is dev_t type; this operation still > deal dev_t as 16 > bits.------------------------------------------------------------------- > ----------------------------------- > > 5. rpmio/rpmrpc.c line 695 > st->st_rdev = ((maj & 0xff) << 8) | (min & 0xffff00ff); > > use major device number and minor device number to construct dev_t, the > result st->st_rdev may look like this: 1111 1111 1111 1111 MMMM MMMM mmmm > mmmm But the correct result should be: > mmmm mmmm mmmm MMMM MMMM MMMM mmmm mmmm > "M" means major device number. > "m" means minor device number. > ----------------------------------------------------------------------- > ------------------------------- > > 6. build/files.c line 1546 > statp->st_rdev = ((fl->devmajor & 0xff) << 8) | (fl->devminor & 0xff); > > this use the similar method as makedev() list befor to get device number. > so it has the same > problem.---------------------------------------------------------------- > -------------------------------------- > > 7. lib/fsm.c line 704 > dev_t finalRdev = (fi->frdevs ? fi->frdevs[i] : 0); > This assign fi->frdevs[i] to finalRdev if fi->frdevs isn't NULL. but > fi->frdevs[i] is unsigned short type, 16 bits and finalRdev is dev_t > type, 32 bits; so it may cause > problem.---------------------------------------------------------------- > -------------------------------------- > > Since I didn__t find any information about this aspect in homepage of > this package, I wonder whether the latest version has completed the > device number extension? If not, will it be completed in the future? And > when? > > Looking forward to answering. > > > Best Regards > -------------------------------------------------- > Zhou Wei > Dept. of Technology and Development > Nanjing Fujitsu Nanda Software Tech. Co., Ltd.(FNST) > No. 16-5, Guanzhou Rd., Nanjing, P.R.China > PHONE: +86+25-6630523-633 > FUJITSU INTERNAL: 79955-633 > FAX: +86+25-3317685 > Mail: zhouwei@xxxxxxxxxxxxxxxx > -------------------------------------------------- -- Clean custom Red Hat Linux rpm packages : http://freshrpms.net/ Fedora Core release 1 (Yarrow) - Linux kernel 2.4.22-1.2163.nptl Load : 0.10 0.23 0.26 _______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
